The Trump administration is finalizing a voluntary framework that would establish a standardized process for AI companies to submit their most advanced models for government review before public release. A draft was circulated between July 13-14, 2026 by the Office of the National Cyber Director to OpenAI, Anthropic, and Google, which provided joint edits to the document. The framework aims to replace the current ad-hoc national security reviews of advanced AI models with a more predictable and structured approach. It is explicitly voluntary, with no mandatory licensing or preclearance requirements. Companies remain cautious about full commitment until testing rules are clearly defined, particularly regarding frontier model definitions and open-source treatment, which could impact decentralized AI protocols.
Source: Read the original article
